> For the complete documentation index, see [llms.txt](https://docs.rubyscore.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.rubyscore.io/how-rubyscore-works/mathematical-model.md).

# Mathematical Model

**RubyScore** transforms a wallet’s on-chain behavior into reputation signals using a **multi-factor model** that includes:

* **Metric normalization** (per chain and over time),
* **Cross-chain aggregation**,
* **Component computation** for **quality / consistency / diversity** (with anti-fraud controls).

### Core signals (non-exhaustive)

* **Amount on balance** — current and averaged balance.
* **Gas spent** — total gas burned.
* **Transactions with unique contracts** — count of distinct contracts; diversification by category (DeFi / NFT / Bridge / DAO).
* **Transactions on different days** — active days.
* **Transactions on different weeks** — weekly distribution of activity.
* **Transactions on different months** — month-over-month activity stability.
* **Transaction volume** — notional volume, normalized per chain/category.
* **Number of transactions** — total count and call typology.

**Additionally considered:**

* **Depth of action sequences** (e.g., bridge → deposit → swap → LP → governance),
* **Economic rationality** (fee/volume ratios, asset retention),
* **Temporal stability** (regularity vs. one-off spikes),
* **Pattern uniqueness** (deviation from mass scripted patterns).

### Scoring approach

RubyScore uses a **proprietary scoring model**: a composition of **weighted, normalized metrics** with **anti-fraud rules**. The output is calibrated to a unified scale and can be tuned for a specific product or ecosystem.

**The final score indicates how closely a wallet’s behavior matches the ideal profile of a valuable audience for a given task or network.**

This approach makes reputation **verifiable, portable, and configurable**—the foundation for trust and fair incentives in Web3.